https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123651
--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
So we'r diagnosing test () with just
use (SR.5_3(D));
I have a patch to disregard DECL_ARTIFICIAL variables as if there were no
associated decls with the SSA name. That makes the reporting consistent.
Eventually we could leverage yet another wording if there's a meaningful
location instead, as to say
t.ii:5:19: warning: uninitialized use of variable [-Wuninitialized]
5 | ~Count() { use(p); }
| ~~~^~~
but I'm sure this will add quite a bit to the amount of diagnosed cases
and given we are using the 'context' location (here that's the call),
there'd be quite some ambiguity with respect to the actually uninitialized
argument, if there are multiple ones.
